Introduction
Step into the realm of NBA Analytics, where the intersection of data and the basketball court creates an exhilarating blend of sports and statistical insights. In this age of technological progress, basketball has transcended the confines of the court, with teams and analysts leveraging data to secure a competitive advantage. Over the last few decades, basketball has become one of the world’s largest sports. Names like Jordan, Kobe, and LeBron now have worldwide recognition, and children of diverse cultures now grow up idolizing their favorite players and wearing their jerseys. This has become evident with the now-commonplace presence of non-American players within the NBA, America’s basketball league, and the world’s most elite.
For those unfamiliar with basketball, the sport involves two teams of five players who compete to pass a single ball through their respective hoops. A successful shot from close range awards a team two points, a shot from long-range (beyond the three-point arc) awards a team three points, and a penalty shot called a “free throw”, (given by the officials for a foul) awards a team one point. As the sport has developed, many established methods for scoring have emerged: layups, slam dunks, mid-range jumpers, alley-oops, floaters, finger rolls, hook-shots, three-point shots, etc.. Professional teams spend millions of dollars on research to increase their scoring efficiency. Also, similar amounts are spent in an effort to decrease the opposing team’s scoring efficiency, i.e. optimizing defensive schemes. With the recent rapid increase in computer power and the development of advanced methods, these practices have created a permanent data-centered community within the sport.
The ultimate goal, obviously, is to win the NBA championship. Only one of the thirty teams can win each year, and many organizations have never brought home the trophy throughout the league’s fifty-year existence. By contrast, there are some teams with multiple championships and some with dynastic periods of consecutive championships.
These discrepancies are the primary motivating factor for the sport’s turn towards a data-centric development approach: any edge that a less-successful team can secure may be the deciding factor that allows it to win a championship. One example of data fuelling a change in the game is the rise of the three-point shot and the subsequent rise of the Golden State Warriors from one of the league’s worst teams to the dynastic juggernaut they are today.
Once dominated by the big men of the ’90s and early oughts, today’s game is primarily focused on distance shooting: the 3-point shot that awards a team an additional point. Contrary to the intuition of previous generations, many of the league’s tallest players are now valued for their shooting ability over their physical superiority in the paint (the area most proximal to the hoop). There has been much debate in recent years over whether this shift has degraded the sport. (One often hears the refrain “Steph broke basketball”, referring to the outstanding 3-point shooting ability of multi-MVP-award-winner Stephen Curry of the Golden State Warriors.) Thankfully, basketball is one of the most quantitatively documented sports, so we can turn to recorded data to shed some light on this debate as well.
The goal of this project is to explore how statistical theory can be applied to the vast amount of NBA data at our fingertips. Also, it is a good practice to get a better understanding and feel of the data, in order to then explore the more nuanced topics within the chosen topic. For example, the plot below visualizes the possible momentum effect of successfully performing a slam dunk. Dunks are entertaining to watch, and often seem to impart a boost in momentum for the dunker’s team. But does the data substantiate that boost? This, along with much more, are areas of interest within the project.